Phenotype-driven genetics in the mouse has long been understood to be a powerful method for linking human genes with their functions and will contribute significantly to the annotation of the human genome. However, animal husbandry requirements and the difficulty of many screening regimens limit the efficiency of standard in vivo methods. We propose that a novel combination of recent technological advances, chemical mutagenesis in mouse embryonic stem cells and expression array screening, would improve vastly the throughput of this robust genetic approach. Large numbers of mutants could be efficiently produced and screened for thousands of expression phenotypes prior to making mutant mice for biological studies of the mutated genes. We propose to test this novel approach using a set of mutant ES lines heterozygous for mutations in genes with well-described functions. The specific aims of this R21 proposal are 1) Determine the power of expression profiling to distinguish among mutant ES cell lines and 2) Compare the effects of a mutation on expression profile in ES cells and differentiated cell types.